Part of the wave is cut

[ichristophorus]

I have mp3 (256b) where on the start there is sound of a very low level. When this file is used to generate wave part of the wave (in this case 0.07s is missing). When the same mp3 is rewritten on this way that the music is true silent in that region the generated wave is excellent (comparing to other tools).

[chrisn]

I'd like to take a look at the MP3 file, if you're able to share it with me (either here or by email if you prefer).

[ichristophorus]

Hi,

We made comparison using Audocity and peaks.js on some other tracks all is fine, but on this track real start of the music on audocity is about 0.10 second on peaks.js about 0.04 second. To be sure that all is fine we took peakjs demo with sonata using both tools and with this track result is exactly the same with precision 1/100 of second. I know that audiowave app is based on the audocity (it is written in doc) so both results should be the same. I will appreciate the support very well.

Thank you,

Krzys

track: https://s3-eu-west-1.amazonaws.com/tmm-content/B/B446011.tmm.spdc.mp3

https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail Wolny od wirusów. www.avast.com https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

On Wed, 23 Jan 2019 at 16:28, Chris Needham notifications@github.com wrote:

I'd like to take a look at the MP3 file, if you're able to share it with me (either here or by email if you prefer).

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/bbc/audiowaveform/issues/85#issuecomment-456844866, or mute the thread https://github.com/notifications/unsubscribe-auth/AKPyNRLYfwKlntaOpQKKPBWotRvLKGQuks5vGH-vgaJpZM4aOpPZ .

[chrisn]

This is likely related to https://github.com/bbc/audiowaveform/issues/49. Your MP3 file has an information frame at the beginning, which audiowaveform skips over, and this could explain the difference you see at the start of the audio. This blog post goes into the detail.

[ichristophorus]

But if audio waveform is based on the audocity why the logic is different? and it means that processing of the wave in my opinion is improper. Of course human ear is not able to find difference of 0.07s in time but the question remains. Thanks for the explanation.

On Wed, 23 Jan 2019 at 22:20, Chris Needham notifications@github.com wrote:

This is likely related to #49 https://github.com/bbc/audiowaveform/issues/49. Your MP3 file has ID3 tags in it, which audiowaveform skips over, and this could explain the difference you see at the start of the audio. This blog post https://thebreakfastpost.com/2016/11/26/mp3-decoding-with-the-mad-library-weve-all-been-doing-it-wrong/ goes into the detail.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/bbc/audiowaveform/issues/85#issuecomment-456970627, or mute the thread https://github.com/notifications/unsubscribe-auth/AKPyNcZ82iUa4Q8lpTX7Kqj2WxfkIfxJks5vGNI7gaJpZM4aOpPZ .

[chrisn]

audiowaveform does not use the same MP3 decoding as Audacity, so I think the question should be: does the visualisation agree with the audio you hear, in Peaks.js? Matching Audacity behaviour is not the goal.

[ichristophorus]

You know this is hard to hear difference of 0.07s. But from listening point of view yes.

On Thu, 24 Jan 2019 at 11:09, Chris Needham notifications@github.com wrote:

audiowaveform does not use the same MP3 decoding as Audacity, so I think the question should be: does the visualisation agree with the audio you hear, in Peaks.js? Matching Audacity behaviour is not the goal.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/bbc/audiowaveform/issues/85#issuecomment-457141201, or mute the thread https://github.com/notifications/unsubscribe-auth/AKPyNbQGC5_-i-N6ydM_FBMJapz0VvLIks5vGYZegaJpZM4aOpPZ .